Summary: | Multi-robot systems are of great importance for tasks that require more robust and faster operation. It can be applied to different scenarios such as search and rescue, modern agriculture, self-driving vehicles and more. The principle work is accurately locating each robot’s position in the global coordinate frame, so that the map generated by each robot can be continuously fused together. In this paper, an approach on helping the principal researchers for collecting and pre-processing image data has been done. Furthermore, the author uses a convolutional neural network (CNN) architecture to train the model and correctly recognize locations under dynamic environment.
|